摘要
为提高评价模型的计算精度,本文利用投影寻踪、蚁群算法、插值型曲线和水环境质量评价标准,为水环境质量综合评价建立了一种新的模型——蚁群插值模型.该模型可依据样本自身的数据特性寻求最佳投影方向.通过最佳投影向量和评价指标向量的内积把水环境质量多维评价样本指标综合成一维投影指标.根据该投影指标值和对应等级的分布,可建立水环境质量评价的蚁群分段插值模型(ACSIM).文中给出了新模型实施的详细步骤,并在各等级中产生建模样本500个,利用该模型对长江流域19个主要站点的水环境质量进行了综合评价.ACSIM评价结果介于Bayes判别分析方法和主成分分析方法之间,且精度较高.ACSIM的评价结果是实数值,可自动优化指标权重,具有直观、有效和通用性,可应用于环境质量综合评价问题.
In order to increase precision of water environmental quality comprehensive assessment, a new model, ant colony subsection interpolation model, or ACSIM, is presented for water environmental quality assessment with projection pursuit, ant colony algorithm, interpolation shape curve and assessment standards of water environmental quality. This model seeks optimum projection direction according to sample data characteristic. Multidimensional indexes of water environmental quality can be synthesized with one dimension projection index by inter-product between optimum projection vector and assessment index vector. The interpolation model for assessing water environmental quality is established by distributions between index value and corresponding grades. Samples can be naturally assessed according to projection indexes, and a detailed step is developed for this model. Five hundred samples in each grade were adopted to test parameter stability in the model. This model was used to assess water resources reproducible ability of 19 administrative divisions of Changjiang River. Compared with Bayes and PCA models, ACSIM had high precision and good applications. ACSIM was also used to design weights of an index system. It is concluded that ACSIM is a superior assessment method for environmental quality comprehensive assessment.
出处
《北京师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第4期429-433,共5页
Journal of Beijing Normal University(Natural Science)
基金
国家科技支撑计划资助项目(2006BAB04A09)
国家重点基础研究发展规划资助项目(G2003CB415204)
水文水资源与水利工程科学国家重点实验室开放研究基金资助项目(2006411411)
关键词
水环境质量
综合评价
蚁群算法
插值模型
water environmental quality
comprehensive assessment
ant colony algorithm
interpolationmodel
